MLLM Is a Strong Reranker: Advancing Multimodal Retrieval-augmented Generation via Knowledge-enhanced Reranking and Noise-injected Training.
Zhanpeng ChenChengjin XuYiyan QiJian GuoPublished in: CoRR (2024)
Keyphrases
- domain knowledge
- knowledge acquisition
- information retrieval
- knowledge base
- video search
- retrieval method
- noise level
- retrieval systems
- relevance feedback
- supervised learning
- denoising
- image database
- image classification
- knowledge representation
- visual features
- domain experts
- semantic information
- image retrieval
- noisy data
- training set
- expert systems
- multimedia